Mitsubishi Research Institute, Inc. published a report on 13 October recommending an A2A economy in which organizations autonomously conduct transactions across organizational boundaries, assuming the widespread adoption of generative AI and AI agents. It also sets out measures for Japan to take an active role in shaping the A2A economy.
The report outlines the background and objectives, depicting a future in which AI agents evolve from providing support for individual tasks to coordinating with multiple agents, enabling transactions across organizational boundaries to be automated. This is expected to create new markets and reshape industrial structures, and the report points out that the focus will be on how the Japanese government and companies participate in establishing the rules and ecosystem.
As an overview of the report, it presents a definition of the A2A economy and explains that automating transactions in which multiple agents coordinate across organizational boundaries could transform the foundations of economic activity. Initial use cases include manufacturing supply chains, while future applications are expected to include intermediary fields such as real estate, finance and human resources, as well as more advanced social infrastructure operations. The report also identifies risks specific to multi-agent environments, including erroneous decisions, collusion and conflicts, along with the challenge of the costs required to ensure reliability.
To enable Japan to participate in the formation of the A2A economy, the report considers it important to develop functions that ensure trust as shared platforms and recommends building industry-specific middle platforms. It states that standardizing industry-specific transaction functions would contain costs, and that the development of digital infrastructure and the strengthening of governance should be advanced. In terms of policy, the government is expected to support greater legal predictability and the creation of initial demand.
Looking ahead, MRI states that it will continue its research and recommendations on the industrial implementation of A2A, the division of roles between humans and AI agents, rules and governance, and the future of digital infrastructure, taking into account challenges in both technology and institutions.